<p>Natural hazards and environmental risks, including air pollution, pose significant threats to human health and urban sustainability, highlighting the need for spatially explicit analytical approaches to support effective monitoring and management. This study investigates the spatiotemporal distribution of PM₁₀ (particulate matter ≤ 10&#xa0;μm) concentrations in the Eastern Marmara Basin, Türkiye, using datasets from 2014, 2020, and 2024 within a Geographic Information Systems (GIS)-based framework. Meteorological, land-use, and topographic variables were integrated into a spatial database to support analysis. Geographically Weighted Regression (GWR) was used to examine spatially varying relationships between PM₁₀ concentrations and environmental factors. The results indicate that meteorological variables are the primary drivers of PM₁₀ variability, whereas land use and topography have moderate but locally significant effects. Local R² values reveal pronounced spatial heterogeneity and highlight areas where pollutant levels are highly sensitive to environmental and anthropogenic influences. The findings demonstrate the effectiveness of combining GIS and GWR to capture spatial non-stationarity in air pollution dynamics and to provide a reproducible framework for environmental risk assessment and spatially informed decision-making.</p>